Ameliorative medical AI: feminist hermeneutics and algorithmic normativity in healthcare
摘要
As healthcare systems prepare to implement AI-driven clinical decision support systems (CDSS), philosophical debate has increasingly examined the algorithmic normativity shaping the clinical encounter. Early discussions often framed AI as displacing physicians, while more recent work, inspired by medical hermeneutics, presents CDSS as dialogical ‘partners’, expanding interpretive resources within deliberative models of care. While a deliberative model presents a more accurate representation of the exact nature of the clinical encounter, opting for a dialogical model for AI implementation without attention to power risks obscuring how structural inequalities shape which voices are heard and which knowledges are valued. I argue that while a hermeneutical, dialogical framing of the physician–patient–AI triad presents a step forward, it fails to account for the algorithmic and biomedical normativity pervading each axis of this encounter. Drawing on feminist receptions of hermeneutics and the epistemic injustice in healthcare literature, I develop a repoliticized account of clinical dialogue attentive to these dynamics. Through a case study of endometriosis, I illustrate how CDSS may either entrench restrictive biomedical classifications or, if designed from the margins, expand hermeneutical resources and actively ameliorate care inequities. I conclude by sketching the contours of a feminist medical hermeneutics for AI-informed care, proposing “ameliorative medical AI” as a framework for leveraging CDSS as tools for advancing healthcare equity.